Self-learning vector memory for AI agents — single-file .rvf cognitive container with HNSW search, episodic Reflexion memory, causal graph + Cypher, 9 RL algorithms, Thompson Sampling bandit, 41 MCP tools, hybrid (BM25 + dense) retrieval, GNN attention. 1
- capability exposure inferred + 35
- tool safety inferred + 12
- supply-chain attested + 6
- trust mitigators mixed − 8
attested inferred mixed
The A–E grade is our heuristic synthesis — a "review this" prompt, not a verdict. Each factor is tagged by what backs it: attested (a verifiable record), reported (a third party's claim), or inferred (our own heuristic, e.g. permissions). See methodology.
graded 9m ago · see ecosystem CVEs →
- B · 33 → C · 45
- C · 48 → B · 33
- D · 60 → C · 48
No known CVEs for this server.
- high dangerous code
credential logged in 1 file(s)
analyzed v3.0.0-alpha.20 · analyzer v32 · 14h ago
danger signals1
- credential in logs credential in log package/dist/src/cli/agentdb-cli.js :662
log.info(`Clients can connect using: agentdb sync connect <host> ${port} --auth-token ${authToken}`);
- supply-chain +6 supply-chain hub →
[](https://mcpobservatory.com/servers/npm:agentdb/security) Heuristic, inferred signals — false positives (legitimately powerful tools, forks, language ports) are expected. Treat each as "review this", not a verdict. See the ecosystem-wide picture on the security hub, or the fleet security of ruvnet.